A hospital-based case-control study was conducted to determine whether residential exposure to magnetic fields increased risk for canine lymphoma in pet dogs. Cases were patients at a veterinary teaching hospital with histologically confirmed lymphoma diagnosed between 1987 and 1990. Hospital controls with other forms of cancer were obtained by frequency matching on zip code and year of diagnosis. Information regarding the dog's activity patterns, residence history, and exposure to potential confounders was obtained by telephone interview. Wire codes and magnetic fields were measured at the homes at diagnosis of 93 cases and 137 controls. When exposure was categorized into two levels (high or very high wire codes compared with low, very low, or buried lines), the risk was elevated (odds ratio (OR) = 1.6, 95% confidence interval (CI) 0.9-2.9) and increased (OR = 1.8, 95% CI 0.9-3.4) after adjustment for potential confounders. Dogs that lived in homes with very high current codes had the highest risk (OR = 6.8, 95% CI 1.6-28.5). Moderate, imprecise increases in risk (odds ratios of 1.5-1.9) were found for residence in a home with a sidewalk (plumbing), backyard, or front yard magnetic field of 2.0 mG or greater, but not for indoor measurements at this level. Risk increased among dogs that spent more than 25% of the day outdoors. Laboratory and observational studies of dogs as an animal model for the effects of magnetic fields are recommended.